This is an archived copy of the Xen.org mailing list, which we have preserved to ensure that existing links to archives are not broken. The live archive, which contains the latest emails, can be found at http://lists.xen.org/
Home Products Support Community News


[Xen-ia64-devel] [PATCH 0/16] make xenheap configurable unpinning xenhea

To: xen-ia64-devel@xxxxxxxxxxxxxxxxxxx
Subject: [Xen-ia64-devel] [PATCH 0/16] make xenheap configurable unpinning xenheap
From: Isaku Yamahata <yamahata@xxxxxxxxxxxxx>
Date: Mon, 7 Jan 2008 16:52:33 +0900
Cc: Isaku Yamahata <yamahata@xxxxxxxxxxxxx>
Delivery-date: Sun, 06 Jan 2008 23:54:35 -0800
Envelope-to: www-data@xxxxxxxxxxxxxxxxxx
List-help: <mailto:xen-ia64-devel-request@lists.xensource.com?subject=help>
List-id: Discussion of the ia64 port of Xen <xen-ia64-devel.lists.xensource.com>
List-post: <mailto:xen-ia64-devel@lists.xensource.com>
List-subscribe: <http://lists.xensource.com/cgi-bin/mailman/listinfo/xen-ia64-devel>, <mailto:xen-ia64-devel-request@lists.xensource.com?subject=subscribe>
List-unsubscribe: <http://lists.xensource.com/cgi-bin/mailman/listinfo/xen-ia64-devel>, <mailto:xen-ia64-devel-request@lists.xensource.com?subject=unsubscribe>
Sender: xen-ia64-devel-bounces@xxxxxxxxxxxxxxxxxxx
User-agent: Mutt/
These patches are intended for post xen 3.2 and probably moot for performance.
This patchset depends on the patchset which fixes xen relocation 
I sent out last year. I went further to remove pinning down xenheap.

Currently Xen/IA64 have to pin down xenheap with DTR so that its size
must be one of TLB mappable sizes and can't be changed easily.
This is the scalability restriction which should be removed.
The patchset removes it and all pages allocation in the ia64 specific code
are done from domheap. Thus xenheap size is configurable up to 4GB.
(4GB restriction is caused by __pickle()/__unpickle())
However this change potentially would increase tlb misses and decrease
the performance. 
Presumably benchmarks should be done to confirm that the performance degrade
is acceptable.

This patchset does
- make hyperprivops.S/optfault.S refer only to struct vcpu.
  If necessary, copy the member of struct arch_domain to struct arch_vcpu.
- allocate vm buffer for VT-i before boot allocater to be pinned with
- pin down the VPD area with DTR and ITR.
- don't pin down xenheap.
- allocate pages from domheap instead of xenheap.
- add xenheap_megabytes boot option.
- some bug fixes and clean ups.


Xen-ia64-devel mailing list